At first glance, using the first_visit event to identify new users seems straightforward. Filter by event name, count distinct user_pseudo_id, and call it a day.
But relying on this signal introduces several problems:
• A “first visit” reflects the device, not the person. A single individual visiting from multiple devices generates multiple first_visit events unless you use a unified User-ID system.
• The event may be delayed or delivered out of order, especially in mobile environments or when offline activity syncs later.
• It captures an interaction, not a verified new user. Treating every first_visit event as a new user inflates your numbers.
Better methods include:
• Using User-ID when possible and deduplicating across devices
• For setups without cross-device stitching: identifying users whose first recorded session (ga_session_number = 1) occurs within the target period
• Leveraging user_first_touch_timestamp for a more dependable “first seen” indicator
New users require validation, not assumptions. GA4 gives the clues, but BigQuery helps you establish the truth.
Want to get all my top Linkedin content? I regularly upload it to one Notion doc.
Go here to download it for FREE.
